When methods in computational mechanics leave the laboratory and enter industrial workflows, the decisive question is no longer theoretical convergence but wall-clock time: how quickly can a software deliver a result of prescribed accuracy? In practice, however, performance figures are still reported in incompatible ways—if at all—making it almost impossible to compare, say, different finite-element flavours, AI surrogates, GPU kernels or emerging quantum solvers on equal footing. The situation is further aggravated for transient problems: while much of today’s research publicises ever-larger or complex simulations on heroic HPC machines, genuinely time-dependent analyses at the scaling limit receive far less attention, and progress in their time-to-solution is rarely documented. This minisymposium therefore launches a community-wide “time-to-solution” challenge that puts absolute speed for transient problems centre stage while remaining method-agnostic and scientifically rigorous.